Medical Hypotheses 17: 271-275, 1985

INTERPERSONAL DISTANCE:A FACTOR IN INFECTIOUS DISEASE MORBIDITY AND MORTALITY?

T.A. Gerace and R. Vorp. Department of Epidemiology and Public Health(R-669), University of Miami School of Medicine, P.O. Box 016069, Miami,FL 33101.

ABSTRACT

We hypothesized a greater incidence of diseases transmitted by airbornevectors in countries where people interact at shorter distances from eachother compared to countries where people interact at greater distances.Therefore, we examined the relationship between the interpersonal dis-tance of pairs of people from different countries and the incidence ofinfectious disease in those countries. Morbidity and mortality data for1981 were obtained from WHO reports, while interpersonal distance esti-mates were taken from Watson's study of proxemic behavior. Respiratorytuberculosis morbidity was negatively correlated with interpersonal dis-tance (r = -.51, p = .037), while respiratory tuberculosis mortality wasnot (r = -.42, p = .051). Mortality from non-respiratory tuberculosiswas negatively related to interpersonal distance (r = -.60, p = .007).To verify these correlations, we repeated the analyses usinq incidencedata from 1977. Neither of the correlations remained significant, al-though for respiratory tuberculosis morbidity r = -.44 (p= .059). Ourinability to replicate the original findings suggests the significantresults for 1981 may have been spurious.

INTRODUCTION

Personal behaviors such as smoking and the consumption of dietary fatshave long been associated with increased risk for chronic diseases,e.g., cancer and corornary heart disease (1). The role of humanbehavior in infectious disease epidemiology has been less often studiedalthough attention has been recently given to the role of handwashing innosocomial infections (2) and sexual behavior in AIDS (acquired immunedeficiency syndrome) (3).

We hypothesized a greater incidence of infectious disease transmitted byairborne vectors in countries where people interact at shorter distancesfrom each other compared to countries where they interact at greaterdistances. Therefore, we examined the relationship between the interper-sonal distance of pairs of people (4) from different countries and theincidence of disease in those countries. To our knowledge this is

271

the first examination of an association between a social behavior, i.e.,interpersonal distance, and the incidence of infectious diseases, e.g.,respiratory tuberculosis and meningococcal meningitis.

METHODS OF PROCEDURE

Data Sources. Unadjusted (i.e., for sex and age) morbidity and mortal-ity data by country were obtained for 1981 from the World Health Organi-zation (5). Interpersonal distances by country were taken from Watson'sstudy of proxemic behavior (6). Countries were selected for analyses ifdata were available for both the disease of interest and interpersonaldistance (Table l), thus accounting for differences in the number ofpairs available for computing various correlations (ref Table 2).

Table 1. Countries With Interpersonal Distance and Incidence of DiseaseData.

Australia France Japan PeruAustria German Democratic Rep. Kuwait PhilippinesBolivia German Federal Rep. Mexico ScotlandCuba Hong Kong Netherlands ThailandEcuador Iraq Norway TurkeyEl Salvador Italy Paraguay United StatesEngland, Wales Venezuela

Morbidity data were used as the primary endpoints because they are lessaffected by the level of health care in a country than are mortalitydata. Mortality data were also studied since they have the advantage ofbeing better documented and not susceptible to diagnostic error.

Statistical Analyses. Univariate correlations were calculated afterpairing interpersonal distance and the incidence of disease of interestfor each country. Both Pearson coefficients of correlation and Ken-dall's Tau were determined using the Statistical Package for the SocialSciences (7). Kendall's Tau was chosen over Spearman's Rho since Tauhandles ties (e.g., in interpersonal distances) in ranking better (7).The alpha level was set at .05. Negative correlations indicate that asinterpersonal distance decreases the incidence of disease increaser.

Control Correlations. To control for variability in health conditionsand health care across countries, we examined the relationship betweenthe incidence of infant mortality (authors' index of the health carestatus of a nation) and interpersonal distance. We reasoned if inter-personal distance and health care status (i.e., infant mortality) wererelated, one could argue that an observed relationship between interper-sonal distance and the incidence of disease was confounded by the rela-tionship between interpersonal distance and the health care status of anation.

As a control against reporting bias, we examined the relationship be-tween the incidence of syphilis, transmitted by direct physical contact,and interpersonal distance. We did not expect to find a relationshipsince the distances people position themselves from each other in gener-al social situations would not be expected to be related to syphilismorbidity. Therefore, finding a relationship between interpersonal

272

distance and syphilis morbidity would certainly make other observed re-lationships suspect. Examining the relationship to syphilis also indir-ectly tests for generalized susceptiblity to disease, perhaps as <a con-sequence of immunosuppression resulting from a crowded environment.

RESULTS

The Pearson coefficients of correlation are summarized in Table 2. Res-piratory tuberculosis morbidity was negatively correlated with interper-sonal distance (r = -.51, p = .037), while respiratory tuberculosis mor-tality was not (r = -.42, p = .051). Mortality from non-respiratorytuberculosis was negatively related to interpersonal distance (r = -.fXl,p = .007). Neither meningitis mortality (r = .-23, p = .191) normorbidity (r = -.02, p = .457) were related to interpersonal distance.

Table 2. Correlations Between Interpersonal Distance and Incidence ofDisease For Selected Countries.

NumberDisease of Pearson r p value

Countries

Respiratory TuberculosisMortality 16 -.42 .051Morbidity 13 -.51 .037

Non-respiratory TuberculosisMortality -.6D .007Morbidity .Ol .490

Meningococcal MeningitisMortality 16 -.23 .191Morbidity 22 -.02 ,457

Infant Mortality 13 -.24 .219

Syphilis Morbidity 19 -.21 .I96

Similarly, neither infant mortality (r =tality (r =

-.24, p = .219) nor syphilis mor-

distance.-.21, p = .196) were significantly related to interpersonal

These latter two findings mitigate the argument that the signif-icant correlations between interpersonal distance and both respiratory tu-berculosis morbidity and non-respiratory tuberculosis mortality could beexplained by level of health care, reporting bias, or general suppressionof the immune system,

In order to verify the significant correlations, we did the following.First, we repeated the analyses using incidence data from 1977 (8). Neith-er of the statistically significant tuberculosis correlations remained sig-nificant, although the correlation for respiratory tuberculosis morbiditywas -.44 (p = .059). Second, because of the small sample sizes we repeaterthe correlational analyses using a non-parametric statistic, Kendall's TauRespiratory tuberculosis morbidity was not related to interpersonal dis-tance in 1981 (Tau = -.33, p = .141) or 1977 (Tau = -.26, p = .231).Neither was respiratory mortality (1981, Tau = -.19, p = .348; 1977, Tau =.23, p = .175). Non-respiratory tuberculosis mortality remained signifi-cant for 1981 (Tau = -.45, p = .028), but was not significant for 1977 (Tat= -.14, p = .444).

273

CONCLUSIONS

Our inability to replicate the original findings suggests the significantresults from 1981 may have been spurious. A true relationship betweeninterpersonal distance and the incidence of disease would be expected toremain stable over several years. The overall results for respiratorytuberculosis morbidity, however, indicate further research is needed inthis area.

Future studies might benefit from examining separately males and femalessince females tend to interact at smaller distances from each other thanmales (9) and in some cultures interact less often with each other thanmales. Furthermore, since the incidence of respiratory tuberculosis variesby sex and age, subsequent studies should control these factors. Not beingable to adjust the incidence of disease data for the sex and age distribu-tions within each country may have reduced the power of this study todetect significant correlations between interpersonal distance and theincidence of infectious disease.

Although our initial findings were not supported after rigorous scruti-ny, this study illustrates several important points. First, new areaswithin behavioral medicine can be examined efficiently by using datathat have already been collected and are easily obtainable. Second, al-though relationships may be found between psychosocial variables and dis-ease, these must be verified using separate data sets whenever possible.Third, special care must be taken in interpreting correlations derivedfrom small sample sizes. Finally, the current study may have been limitedby the representativeness of the data on interpersonal distance. Thesedata pertained to the behavior of foreign students attending college in theUnited States.

1.

2.

3. .

4.

5.

6.

7.

REFERENCES

Gerace TA, Vorp, R. Epidemiology and behavior. In N Schneidermanand J Tapp (Eds), Rehavioral Medicine: A biopsychosocial approach.Hillsdale, NJ, Erlbaum, in press, 1984.

Albert RK, Condie F. Hand washing patterns in medical intensive careunits. New England Journal of Medicine, 304(24):1465, 1981.

Jaffe HW, Bregman DJ, Selik RM. Acquired immune deficiency syndromein the United States: The first 1,000 cases. Journal of InfectiousDiseases, 148(2):339, 1983.

Evans GW, Howard, RB. Personal space. Psychological Bulletin,80(4):334, 1973.

World Health Statistics Annual, Volume I - Vital Statistics and Causesof Death, World Health Organization, Switzerland, 1981.

Watson OM. Proxemic Behavior; A cross-cultural study. Mouton andCompany, The Hague, 1970.

Nie NH, Hull CH, Jenkins JG, Steinbrenner K, Bent DH. StatisticalPackage for the Social Sciences, 2nd Edition. McGraw-Hill, New York,1975.

274

8. World Health Statistics Annual, Volume 1 - Vital Statistics an-l Causesof Death, World Health Organization, Switzerland, 1977.

9. Lett EE, Clark W, Altman I. A propositional inventory of researchon interpersonal distance (Research Report No.1) Bethesda, MD, NavalMedical Research Institute, 1969.

275